Efficient automated translation of procedures in constraint-based language

Abstract

A method is provided for translating sets of constraint declarations to imperative code sequences based on defining an instantiatable object per set, inserting calls to a notification callback mechanism on state modification and defining calls in the constraint context as imperative code sequences that, in response to these callbacks, take actions to maintain these constraints. This notification and callback mechanism can also be employed in connection with external events, thereby providing for efficient implementation of event-sequenced imperative procedures in a constraint programming language.

Description

FIELD OF THE INVENTION[0001]This invention relates to automatic translation of computer languages.BACKGROUND[0002]In the development of computer technology, various programming paradigms have been developed. Two of these paradigms are imperative programming and declarative programming. Imperative programming entails providing step by step imperative instructions (e.g., set a equal to 2, increment c, call procedure f( )) which are combined by a programmer to solve the problem of interest. Declarative programming entails providing top level information (e.g., a set of logical constraints to be enforced), which can often be specified in a manner that more transparently relates to the problem of interest. Imperative programming tends to provide improved execution performance, but imperative programs can be very complex and difficult to develop and maintain. Declarative programs tend to be relatively easy to develop and maintain, but declarative programming also tends to provide relative...

AI Technical Summary

Benefits of technology

[0012]The main capability provided by embodiments of the present invention differs from the example of FIG. 1 in two significant respects. The first of these is to specialize to the case of constraints as the declarative statements. This removes the above-described conceptual difficulty, because a constraint is a purely declarative construct. The second of these is to significantly expand the automatic imperative code generation capability to accommodate situations where constraints are in different contexts and / or different program units from their corresponding data members.

[0082]More specifically, an input source program can be expressed in a constraint programming language that also provides for specification of event-sequenced imperative procedures. The input source program can be automatically translated to an intermediate imperative representation, where notification and callback code for constraints in the intermediate representation provides efficient implementation of constraints, and where notification and callback code for events in the intermediate representation provides efficient implementation of event-sequenced imperative procedures. This imperative intermediate representation can be automatically translated to a computer executable imperative output representation, which in turn can be provided as an output (e.g., to a user).

Problems solved by technology

In particular, when the declarative statement is in the same program unit and same context as the input being changed, as shown, then code to perform A can be automatically inserted by the compiler immediately after every line that alters X. However, there are troublesome conceptual aspects of this construct.

Therefore, such constructs have the potential to complicate software development.

Images